MESSAGE 7 NOVEMBER, AI PROGRAMMING COLLABORATION TOOL Amp Yesterday (6 November) announced the launch of Spaces, which will be the first day of the dayWISH TO BUILD AN AI-ORIENTED SMART BODYAgentNew ways of working with a team。
For users, there is no longer a need to create, name and manage multiple Agent lines manually, but simply to talk to one main Agent, which understands the needs, dismantles the tasks, and automatically generates and coordinates other Agent 1AI with the following relevant screenshots:
Ap points out that in the traditional multi-Agent workflow, users often decide to split, assign and track tasks themselves, and Spaces entrusts this layer of organization to Agent: each Space is the user's dialogue with a single Agent, which starts and coordinates multiple parallel lines for users。
For example, a team could propose “advance a function online” or “conduct a security review” in a Space. The main Agent translates the target into more than one specific task, creating a line for execution and presenting progress in a live card chat。
The Spaces is not just a one-person automation tool, but is designed to be a place where teams work together to plan and direct their work: members can ask questions, feedback or additional ideas in the same chat; each job removed shows the current state and whether manual intervention is required。
Clicking on the taskcard allows the user to enter the corresponding thread to view Agent 's output, continue it iteratively or directly advance delivery. Team members can also comment on the discussion in the news; Amp will read the comments only if @amp is mentioned。
Ap positions Spaces as a “higher” multi-Agent collaboration, where users move from an article-by-article arrangement to a coordinated Agent description of objectives, with the latter responsible for task decomposition, thread creation, progress organization and aggregation of results。
Official examples show that one member has completed the release of service account functions through a single Space; another scenario is a periodic security review. The user needs to click “+” on the Spaces area in the sidebar, set the title and item and explain to Agent the task to be handled。
THIS FUNCTION REFLECTS AI Programming ToolsMoving from “one-person questions and answers” to “teaming, tasking”. When Agent is able to break down autonomously and perform parallel missions, the role of the developer may shift more to defining objectives, reviewing results, addressing key decision-making than manually moving each submission。
Amp is an AI programming Agent and development environment for software development. The core objective is for developers to give Agent what to do, while Agent is responsible for reading codes, changing codes, running tests, iterative functions and carrying out continuous tasks at the cloud or on-line。
Amp can run tools and Shell commands in the user's workspace to check codes, perform tests and quickly invert; it can also assign specific sub-tasks to sub-Agent processing through the Task tool。